Parent issue: benches for quadratic-time VID ADVZ

[ggutoski]

We want benchmark data for the "slow" VID with quadratic-time bottlenecks in dispersal and recovery so that we can quantify the benefit when we eventually eliminate these bottlenecks.

Quadratic bottlenecks

1. dispersal: Multi-eval for Reed-Solomon encoding
2. dispersal: Multi-open for aggregate proofs
3. recovery: Reed-Solomon decoding

Sub issues

• [x] #48
• [x] #49
• [x] #50
• [x] #51
• [x] #52